ICTA’07, April 12-14, Hammamet, Tunisia
An Environment for Deaf Accessibility to Educational Content
Eleni Efthimiou, Stavroula-Evita Fotinea
ILSP – Institute for Language and Speech Processing
[email protected], [email protected]
Abstract
Here it is attempted to present a platform
environment that allows development of various
educational applications fully accessible by deaf users.
Subject to Design for All primes, the environment is
built on methodological principles which adopt sign
language as the basic means for communication of
linguistically uttered educational content. It also
makes extensive use of visual objects to support
comprehension and navigation at all levels of humancomputer
interaction.
Currently
available
instantiations of the environment have incorporated
both video for content presentation and an avatar
based dynamic sign synthesis mechanism. The
educational applications to be referred to when
discussing the design principles and user requirements
taken into account, include one web-based and one offline GSL teaching tool for the same school level (early
primary school) as well as a vocational training tool
for adult users.
1. GSL: an official minority language
Sign languages (SLs) are natural languages
articulated in the 3D space, where the linguistic
massage is organized according to geometrical
parameters for the expression of semantic-syntactic
relations, visually represented on and with the signer’s
body or in the space in front of him/her [21], [22],
[18].
The Greek Sign Language (GSL) has been
developed as a minority non-written language system,
in a socio-linguistic environment similar to those
holding for most other known sign languages. It is a
natural language system used as the mother language
of the Greek deaf community, where estimations raise
GSL natural signers to about 40,600 (1986 survey of
Gallaudet Univ.). In addition to the above, there is also
a large number of hearing non-native signers of GSL,
mainly students of GSL and families of deaf people.
Records of the Hellenic Federation of the Deaf (HFD)
show that in the last five years the demand for classes
of GSL as a second language has radically increased
[12]. This fact finds an explanation in the recently
increased demand for GSL knowledge in the local
language market, given that in the year 2000, GSL was
recognized as an official language of the Hellenic State
(Law 2817/2000), with a direct consequence for the
use of the language in education and official
communication services.
Following the national policy for integration of
people with disabilities in the society, a recent increase
of deaf students in mainstreamed education is to be
noticed. Nonetheless, a considerable proportion of the
deaf student population still remains scattered in other
institutions, minor town units for the deaf and private
tuition. Beyond legislative requirements though,
integration of the Deaf in the society is heavily based
upon the quality of education they receive, and
although GSL is the official language for education of
the Deaf population in Greece, educational material
and tools still remain very poor. This is partly due to
the widely holding misconception that since deaf
people may see, they can access written material.
However, born deaf individuals find it extremely
difficult to make associations between concepts and
written forms. This happens because the written form
of an utterance is a convention for the representation of
sounds, which is incomprehensible in the case where
no perception of sound is possible. According to
statistics of the Hellenic Pedagogical Institute [14], the
average reading capability of deaf adults corresponds
to mid primary school level. This fact seems to be
verified by measurements and estimations from other
sign languages as well [7]. On the other hand, the
nature of the language per se acts as an obstacle
against systematic production of representations of its
linguistic content in huge quantities, if the requirement
is to preserve quality of the delivered message to
native utterance level. This happens because video –
though a rather static, not easily reusable source of
linguistic content– is currently the only representation
means that fully preserves naturalness of the signing
utterance. In the seek of a solution to the limitations
put by the use of video and towards satisfaction of the
requirement for Universal Access [19], continuously
growing in the framework of the Information Society,
the demand for efficient solutions to the problems of
deaf human – computer interaction, also entailing
accessibility to e-content by the deaf, has led to
research for the development of dynamic systems for
the representation of sign language utterances by
means of avatar technologies [5], [6], [13], [15].
This paper focuses on architecture design and
implementation of an educational environment, fully
accessible by deaf users, that allows development of
various educational tools both Internet based and offline.
2. The language barrier in education of the
Deaf
Unrestricted integration to an open society is
heavily based upon quality of education received by
the different groups of population. In other words, one
initial barrier is related to discriminations with respect
to one’s level of literacy.
Most native speakers of minority languages
experience the need to attend school in a linguistic
code other than their mother language. The case of
deaf populations is even more complicate, since born
deaf individuals develop a specific type of bilingualism
[11], [16] deriving from the demand to raise in a
language environment that uses not only
incomprehensible articulation means, but also it
structures concepts and relations of the linguistic
message on completely different grounds than sign
languages do.
From the legislative point of view, inclusion is
supported by incorporation of deaf students either
directly to mainstream education or to special
education units inside mainstream schools or even by
modernizing special schools for the Deaf. In all cases,
offered educational content should be delivered in sign
language. Nowadays, sign language use in classroom
is supported by mature multimedia and digital image
technologies, parallel to the development of especially
user friendly interfaces for end-users who may not be
previously familiar with computer use [1]. Moreover,
appropriate methodologies for content presentation
allow systematic development of tools for the deaf, on
the basis of Design for All principles, where the key
characteristic of such systems is the use of sign
language in order to convey meaning of linguistic
content at any level of interaction with the user.
In this context, teaching of a sign language (GSL in
our case) up to complete mastery becomes crucial, as it
provides the necessary linguistic background for full
academic development of deaf individuals.
In line with the above, we discuss next the main
characteristics of an environment, which supports sign
language teaching as well as teaching in GSL,
incorporating educational material presentation and
testing mechanisms, that make it fully accessible by
deaf end users, pupils as well as tutors.
The environment instantiations used to exhibit
multipurpose implementation of the same basic
architecture, involve a range of educational software
products addressing needs from language teaching in
primary school up to vocational training objects. For
the purposes of discussing architecture principles,
environment ergonomics and content creation
methodology, we make extensive reference to “MTN
1” (from the initials of the full title “I learn the signs”
in Greek) (Figures: 1, 2, 3, 4, 5), the first of a series of
educational software products intended to teach GSL.
The educational goal of this specific show case is to
introduce vocabulary primes at early primary school
level. The same educational goal is served by a web
based prototype system that allows structuring and
presentation of GSL educational material and linguistic
resources, addressing the needs of GSL grammar
teaching to early primary school pupils. In this case,
dynamic sign synthesis through the use of avatar
technologies for 3D sign representation of linguistic
content, is tested as an alternative to the use of video
(Figures: 6, 7, 8) [12], [4], [19].
Figure 1. MTN1 main menu screen.
126
Figure 2. Handshape teaching.
Figure 4. Handshape checking exercise.
Figure 3. Handshape based interactive vocabulary
presentation.
Figure 5. Commentary on exercise execution.
3. Environment design principles and
implementation instantiations
Basic environment architecture is built on general
principles applied to design of educational software
along with the presupposition that linguistic content
has to be represented 3-dimentionally. Image or video
(capture) are the only means for concept clarification,
where decisions on of graphics style and use of
cartoons or photographs depends on age group of users
and educational object. Design modifications, for
example, required for the development of MTN 1,
were organized on the basis of the main educational
goal and user group needs of very young pupils and
their tutors; known through extensive evaluation
procedures of various prototype versions of the
environment.
The use of graphics, bold colours and game-like
screen layouts, as well as incorporation of student
awards for correct responses to lesson exercises and
unit tests derived from design principles applying to
small children. Similarly, user need studies defined
that video ought to be the means of representation for
GSL in the specific case, whereas other basic
functionalities should be adopted, as the need for userdriven repetition of the educationall linguistic message,
as well as the need for navigation details and button
functionalities, presented in GSL and combined with
graphics. The use of graphics in this case follows an
organization with clear semantic structure.
Standard educational means incorporated in this
software for the teaching of vocabulary of GSL are
drawings for visual representation of concepts and
video for the accurate 3D representation of sign
articulation. Video also provides clarifications and
instructions as regards educational goal, execution of
127
exercises and help at all levels of the software.
Educational content is strictly presented by a native
signer tutor. Presentation of all other linguistic
information is provided by the tutor’s assistant, in GSL
also.
The use of drawings for concept representation as
well as the graphical navigation are product specific
design characteristics. The use of drawings and the
respective complete absence of photographs, in this
case, intend in relating a concept with the wider
possible range of objects this may find in real world.
The rational that led to presentation of all navigation
functions with graphics is that graphical navigation has
proven worldwide to be a popular way for children
interaction with computer environments. A general
design characteristic is the strict absence of written
language for presentation of educational content
(except for cases of multilingual terminology as in
Figure 9). Written language may appear only after
selection of the relevant functionality, for use by
hearing tutors, when viewing lists such as tables of
contents.
The educational software MTN 1 reflects the
structure of GSL with respect to principles, features
and rules applying to the vocabulary of the language
[2] while at the same time complies with
methodological principles, thematical structure and
educational activities proposed in the Analytical
Curriculum for the teaching of GSL, set by the
Hellenic Pedagogical Institute.
The MTN platform utilizes GSL as the sole means
of communication, excluding the use of written Greek
in all levels of communication with the students. MTN
supports personalized teaching of the language,
adoptable to individual learner needs, to guarantee
efficient comprehension of the teaching subject by the
student, especially in classes where students present
different levels of command of the language.
Implementation of the MTN educational platform
was based on the available architecture [1], [19] and on
selection of appropriate educational content with
respect to the age of the students; user needs as
reflected in the Hellenic Pedagogical Institute’s
Analytical Curriculum for GSL (http://www.pischools.gr/special_education/kofosi-a/), as well as the
needs/requirements set by user needs studies of the
specific target group (deaf students / GSL teachers) as
recorded via informal evaluation procedures
undertaken to educational prototypes [4]. In addition,
for the development of the MTN platform, studies
regarding teaching conditions of a sign language and
world-wide creation of educational curricula have also
been taken into account [17]. The Internet-based
prototype that exploits avatar potentials for sign
representation adopts the same architecture with the
goal to supprt distant learning.
Figure 6. Full graphics based navigation with
avatar performed vocabulary presentation.
Figure 7. Handshape based vocabulary
presentation with avatar in frontal view.
Figure 8. Color code matching exercise.
128
4. GSL educational content
Educational software instantiations which are
developed making use of the basic environment
design, derive linguistic content from an open pool of
reusable resources as regards both vocabulary items
and grammar rules of GSL.
Vocabulary resources are organized in a lexical
database for signs (the lexical units of sign languages),
where lexical entries are marked for phonological
composition, grammar behavior and semantic
characteristics [2].
Grammar rules of GSL are coded in the form of a
computational grammar, based on results of basic
research on a representative sign language corpus [3].
In the case of development of vocational training
software, a further parameter of content creation is the
demand for introduction to GSL of the domain
terminology. Creation of new terms in any natural
language follows standard procedures [8], [9], [10].
Moreover, since terminology is usually introduced to
receiver languages from the language in which it was
originally created, terminological lists are multilingual
in the default case. To fulfill this task, the environment
is supported by a methodology for term creation in
GSL [2], which allows trilingual (GSL-Greek-English)
representation of terminology items in terminology
intensive educational applications. In this case,
definitions of terminology concepts are visually
available (window at down right side of the screen
Figures: 9, 10, 11).
The currently available GSL resources have fed a
number of applications, which include, except for the
applications discussed in this paper, a bilingual
dictionary of basic GSL vocabulary, with 3.000
entries, a children dictionary of 500 entries and an
NLP based conversion tool from written Greek to
GSL.
As regards MTN 1, educational content is organized
in 5 chapters with 70 lessons and 600 lemmata, which
teach methodology of sign formation in basic signs,
complex signs, synonyms-antonyms and word
families.
The here presented vocational training software
provides educational content for basic computer skills
(7 ECDL topics) with 350 trilingual terminology items
and 650 demonstrators of term usage, covering all
possible appearances of a term in the related thematic
units.
The here presented environment is subject to
continuous evaluation cycles and optimization, while
the linguistic resources databases for GSL content are
constantly updated.
Figure 9. Trilingual terminology presentation for
computer skills with term search by handshape and
video capture example of term use.
Figure 10. Instructions in GSL, supported by term
presentation and example of use.
Figure 11. On-line help in GSL.
129
5. References
[1] Efthimiou, E., S-Ε. Fotinea, «An adaptation-based
technique on current educational tools to support universal
access: the case of a GSL e-Learning platform», Proc. of the
TeL'04 Workshop on “Technology Enhanced Learning”,
Jean-Pierre Courtiat, Costas Davarakis and Thuerry Villemur
(eds), 22 August, Toulouse, France, Workshop to WCC-2004
(World Computer Congress), 2004, pp.177-186
[2] Efthimiou, E., Fotinea, (2004b). «Multipurpose Design of
Greek Sign Language Resources: A factor towards Universal
Access», Proc. of the 1st International Conference "From
Scientific Computing to Computational Engineering" (ICSCCE), Invited Session on Human Computer Interaction, 8Athens, 10 September, 2004, pp. 820-827.
[3] Efthimiou E, S-E, Fotinea, G. Sapountzaki, “Processing
linguistic data for GSL structure representation”, in Proc. of
the Workshop on the Representation and Processing of Sign
Languages: Lexicographic matters and didactic scenarios,
Satellite Workshop to LREC-06, Genoa, 2006, pp 49-54
[4] Efthimiou, E., S-Ε., Fotinea, G. Sapountzaki, “Eaccessibility to educational content for the deaf”, European
Journal of Open and Distance Learning (EURODL), 2007,
forthcoming.
[5] Fotinea S-E, E. Efthimiou, K. Karpouzis, G. Caridakis,
“Dynamic GSL synthesis to support access to e-content”, in
Proc. of the 3rd International Conference on Universal
Access in Human-Computer Interaction (UAHCI 05). Las
Vegas, Nevada, USA, 2005.
[6] Glauert, J.R.W, ViSiCAST: Sign Language using Virtual
Humans, in International Conference on Assistive
Technology (ICAT 2002), Derby, 2002, pp. 21–33.
[7] Huenerfauth, M, “Spatial representation of classifier
predicates for machine translation into American Sign
Language”, in Proc. of the Workshop on the Representation
and Processing of Signed Languages, 4th International
Conference on Language Resources and Evaluation: LREC 04. Lisbon, Portugal, 2004.
[8] International Standard ISO 1087-1:2000 – «Terminology
work – Vocabulary – Part 1: Theory and application»
[9] International Standard ISO/R 704:1968 – «Naming
principles»
[10] International Standard ISO 704 – «Principles and
methods of terminology».
[11] Israelite, N., C. Ewoldt, R. Hoffmeister, “Bilingual/
bicultural education for deaf and hard-of-hearing students: a
review of the literature on the effects of native sign language
and majority language acquisition”, Ontario Ministry of
Education: Toronto, 2002..
[12] Karpouzis K., G. Caridakis, S-E. Fotinea, E. Efthimiou,
“Educational Resources and Implementation of a Greek Sign
Language Synthesis Architecture”, International Journal
Computers and Education, Elsevier, in print, electronically
available since Sept 05, 2005.
[13] Kennaway, J.R., “Experience with and Requirements for
a Gesture Description Language for Synthetic Animation”, in
Camurri, A., Volpe, G. (Eds.) Gesture-based Communication
in Human-Computer Interaction, LNAI 2915, Springer
Verlag, 2004.
[14] Kourbetis, V., Noima Stin Ekpaidefsi (in Greek),
Hellenic Pedagogical Institute, Athens, 1999.
[15] Marshall, I., E. Safar, “Sign Language Synthesis Using
HPSG”, in Ninth International Conference on Theoretical
and Methodological Issues in Machine Translation (TMI),
Keihanna, Japan, 2002.
[16] Mayberry, R., “First-Language acquisition after
childhood differs from second language acquisition: The case
of American Sign Language”. Journal of Speech and
Hearing Research, 36, pp. 51–68, 1993.
[17] Miller-Nomeland, M., S. Gillespie, S. Deaf Studies
Curriculum Guide, Kendall Demonstration Elementary
School, Pre-College Programs, Gallaudet University,
Washington DC, 1993.
[18] Neidle,C., J. Kegl, D. MacLaughlin, B. Bahan,B. R.G.
Lee, The Syntax of American Sign Language, MIT Press,
2000,.
[19] Sapountzaki G, E. Efthimiou, S-E. Fotinea, “Digital
technology in Greek Sign Language teaching in primary
school curriculum”, in Proc. of Hellenic Conference on
Digital Educational Material: Issues of creation, exploitation
and evaluation. April 6-7, 2006. Volos, in print (in Greek)
[20] Stephanidis C, Editorial, International Journal of
Universal Access in the Information Society, 1 pp 1-3, 2001.
[21] Stokoe, W.C., Sign language structure. (2nd ed.), Silver
Spring MD, Linstok Press, 1978.
[22] Sutton-Spence, R., B. Woll, The Linguistics of British
Sign Language. An Introduction, Cambridge, University
Press, 1999.
130